Accretion onto the Companion of

Eta Carinae

Amit Kashi and Noam Soker

Technion – Israel Institute of Technology, Haifa, ISRAEL

IAU XXVII General Assembly

Joint Discussion 13:

Eta Carinae in the Context of the Most Massive Stars

Rio de Janeiro, 13-14 August 2009

ד"בס

The Purpose of the

Accretion Model• The Accretion Model was introduced to explain

observations along the entire orbit, mainly those close around the spectroscopic event.

• We use the standard parameters of the system and show that near periastron the secondary is very likely to accrete mass from the slow dense wind blown by the primary.

• The condition for accretion (that the accretion radius is large) lasts for several weeks. The exact duration of the accretion phase is sensitive to the winds' properties that can vary from cycle to cycle.

The Spectroscopic Event

Figure by A. Damineli

Colliding Winds Structure

Soker (2005)

Why is accretion favorable

close to periastron passage?

• Racc2 : the secondary accretion radius

• Dg2 : distance from the stagnation point to the secondary

• The ratio

Racc2/Dg2 is a

measure to the

influence of the

secondary

gravity on the

primary wind.

• Along most of

the orbit the

Racc2<< Dg2

• Close to

periastron

Racc2/Dg2 ∼ 0.5.

The Accretion Model (cont.)• The accreted mass shuts down the secondary wind for

several weeks, depending on winds' properties.

• Bondi Hoyle accretion onto the secondary (Soker 2005a).

• X-ray luminosity at minimum, due to the collapse of the shocked secondary wind which is the source of X-ray emission (Akashi et al. 2006).

• Disappearance of He I visible lines due to absorption of ionizing flux by the accreted material, which also causes lowering of secondary’s effective temperature (Soker 2007).

• Near-IR decline due to the pause in creating hot dust in the downstream of the conical shell (Kashi & Soker 2008a).

• An accretion belt around the secondary (Kashi & Soker 2009a).

Bondi – Hoyle Accretion

2

2 2 wind1acc2 2 1

wind1

20.2 AU

30 500 km s

GM M vR

v M

Akashi et al. (2006)

Accretion

Radius

What is the Nature of the

Accretion Disk?

• Periastron passage duration is ~6 days.

• Viscosity time of the disk is

• The accretion disk doesn’t reach equilibrium.

• Shakura-Sunyaev α model for a thin accretion disk cannot be used.

• Geometrically thick Accretion Belt.

pp visct tPeriastron viscosity

passage time

3/2 1/22 22 2

2 2visc Kep

1 130

2 10 20da

0y

3s

s

R R R R R Mt t

C H H H R M

Modeling the Accretion• The effective accretion radius of the secondary depends on

several parameters, in particular on the orbital separation r(θ).

Since the accretion radius is very close to the primary, the

primary’s wind acceleration zone is taken into account as a β-

model with two extreme values:1 and 3.

• We consider the two limits: RLOF, taking the accretion

radius as the Roche lobe equivalent radius, , and

Bondi-Hoyle-Lyttleton (BHL) accretion, where:

11 1,( ) 1

Rv r v

r

( )RLR

2 22wind1 2, 1 2,2

wind1

22( ) ; ( )BHL r

GMR v v v v

v

Why is Accretion Favorable close to PP?

accretion radius

≈ 0.26r

primary

radius

density

profile of the

primary wind

1.162AU

Close to periastron:

the secondary accretion

radius is within the

accelerated (dense) zone

of the primary wind. Kashi & Soker 2009a

stagnation

point

Calculation:

Accretion Onto the Secondary

Modeling the Accretion (cont.)

• To calculate the dependence of the primary wind density on the

azimuth angle and on the distance from the secondary, we slice the

cross section into differential arcs.

• The accreted mass from each arc is calculated separately according to

the density at that point, and it is added to the accreted mass.

Modeling the Accretion (cont.)

• Full RLOF-like accretion cannot

occur because the primary spin and

orbital motion are not synchronized

near periastron passage.

• Therefore, for ∼10 days very close to

periastron passage, the accretion

process will be an hybrid of the BHL

and the RLOF mass transfer

processes.

• At the end of the accretion phase the

accretion will be more of the BHL

type.

• Over all, for the standard wind

parameters the total accreted mass is

0.4−3.3×10−6M⊙, with average value

of Macc ∼ 2×10−6M⊙.

RLOF accretion

5 110 M yr

10

1

6

1

wind accretion(Bondi-Hoyle)

5 110 M yr

Removing the Belt • Because of the long viscosity time and the high mass loss rate, this belt

is destroyed mainly by mass loss rather than accretion on the

secondary.

• We assume that the mass loss rate per unit solid angle from the belt is

as that from the secondary.

• The belt covers a fraction δ of the secondary’s stellar surface (for

example, if this belt extends from the equator to latitudes ±30°,then δ =

0.5). The belt will be blown away during a time:

• If the mass loss process starts ∼60 days after the event starts, then the

recovery ends ∼7 months after the event starts. We identify this

duration with the recovery phase of η Car from the spectroscopic

event.

1 1

acc acc 2belt 6 5 1

2

5 month2 10 y10 0.r 5

M M M

tM M M

The Early Exit from the 2009

Minimum

• Early exit from

the 2009 X-ray

minimum after

four weeks,

instead of ten

weeks as in the

two previous

minima.

Figure by

M. Corcoran

0.25x

x

dL

L

1

1

1

2(1 )

x

x

dL dp

L p

1/2

2 2

1 1

0.45M v

M v

1

1

0.7dp

p

-1

1, 300 850 km sv

Early Exit

The Early Exit from the 2009

Minimum (cont.)• We attribute the early

exit in the last cycle to

the primary wind that

we assume was

somewhat faster and

of lower mass loss rate

than during the two

previous X-ray

minima.

• This results in a much lower mass

accretion rate during the X-ray minimum.

-1

1, 900 km sv

-1

1, 500 km sv

The Early Exit from the 2009

Minimum (cont.)

• The changes which occurred near periastron passage, are

most likely due to the strong tidal interaction.

• Most likely, the tidal interaction amplified a small

internal change in the wind properties, e.g., as might

results from magnetic activity.

• Using fluctuations in mass loss and wind velocity that

are within the range deduced from fluctuations in the X-

ray flux outside the minimum, we can account for the

short duration of the last X-ray minimum.

SummaryThe main findings are:

The secondary accretes ~ 2×10−6M⊙ close to periastron.

This mass possesses enough angular momentum to form a geometrically

thick disk, or a belt, around the secondary.

The viscous time is too long for the establishment of equilibrium, and the

belt must be dissipated as its mass is being blown in the reestablished

secondary wind.

This processes requires about half a year, which we identify with the

recovery phase of η Car from the spectroscopic event.

Mass transfer is an important process in the evolution

of close massive binaries.

The high luminosity and ejected mass of many eruptive

events can be explained by mass transfer, e.g., the

Great Eruption of Eta Carinae.